Data analysis in the chemical laboratory II. The end-point estimation in instrumental titrations by nonlinear regression

The regression model for a two-segments titration curve with a break-point at the end-point is analyzed. Both linear and nonlinear shapes of the titration curve segments are treated. An effective and simple method discriminates which of two segments is linear or curved. The point and interval estimates of the end-point are calculated by the nonlinear least squares of curve fitting technique. The nonlinear regression method is applied to any, linear or nonlinear, type of a two-segments titration curve without excluding any titration points to reach the most probable point estimate of the end-point together with its 100(1-alpha)% confidence interval. An accuracy and precision of the proposed end-point estimation is examined on several instrumental titrations. A sample program in S-Plus(TM) is provided.